| Presentation | 2008-02-22 Strategy of Multistage Pattern Recognition Using a Posteriori Probability and Best-first Search : Application to Address Recognition Tomoyuki HAMAMURA, Takuma AKAGI, Bunpei IRIE, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | For faster recognition, coarser classifiers are often employed for pre-classification forming a multi-stage classification system. For example in address recognition, higher level address elements (e.g. city name) are recognized beforehand in order to reduce the number of candidates for lower level ones (e.g. town name). We applied the best-first search for the recogntion problem viewing it from the standpoint of search theory. In this study, we propose a method for calculating a posteriori probability which is effective for search evaluation score. For the first step, it is shown that a posteriori probability can be broken down to the product of a posteriori probability ratio (the ratio between a priori probability and a posteriori probability), which has been proposed by the authors for generarl case. In the next step, more specific method for calculating a posteriori probability ratio is introduced for word recognition problem. The effectiveness of the proposed method is demonstrated by address recognition experiment. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Best-first search / "A posteriori" probability / "A posteriori" probability ratio / Multistage / Pattern recognition / Address recognition |
| Paper # | PRMU2007-243 |
| Date of Issue |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2008/2/14(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Pattern Recognition and Media Understanding (PRMU) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Strategy of Multistage Pattern Recognition Using a Posteriori Probability and Best-first Search : Application to Address Recognition |
| Sub Title (in English) | |
| Keyword(1) | Best-first search |
| Keyword(2) | "A posteriori" probability |
| Keyword(3) | "A posteriori" probability ratio |
| Keyword(4) | Multistage |
| Keyword(5) | Pattern recognition |
| Keyword(6) | Address recognition |
| 1st Author's Name | Tomoyuki HAMAMURA |
| 1st Author's Affiliation | TOSHIBA Corporation() |
| 2nd Author's Name | Takuma AKAGI |
| 2nd Author's Affiliation | TOSHIBA Corporation |
| 3rd Author's Name | Bunpei IRIE |
| 3rd Author's Affiliation | TOSHIBA Corporation |
| Date | 2008-02-22 |
| Paper # | PRMU2007-243 |
| Volume (vol) | vol.107 |
| Number (no) | 491 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |